Further Development in Nondestructive Methods to Gauge Life Expectancy in Ferromagnetic Components

David Millard, Washington State University
Pradeep Ramuhalli, Pacific Northwest National Laboratory
Jacob Fricke, Pacific Northwest National Laboratory

Abstract

Manufactured nuclear components under stresses induced through normal operations cause mechanical fatigue and strain. Depending on their magnitude and distribution they can contribute to increasing the expected life of a component or for its premature failure. Using Barkhausen noise we can analyze the microstructural characteristics without damaging the sample through magnetization or acoustics. The samples in our case are ferromagnetic metals, also known as ferrous metals, from heat treated and rolled steel. A Rollscan 300 instrument and Microscan 600 software were used to acquire Barkhausen noise data from fatigued steel samples. MATLAB software and R software were used to evaluate results of the Microscan 600 to better understand the signal processing algorithms. In order to find a correlation we used a two random variable probability distribution function (PDF). plot We found the difference between the three positions taken on the given sample at each strain level, and with a 95% confidence level we created a plot of data points that found a loose correlation in the data results between both perpendicular and parallel testing. Using these results we can compare older sets of data and create an accurate prediction of stress levels induced upon nuclear components. We hope to create more precise predictions in the near future using alternative methods, such as statistical calibration techniques to find closer one‐to‐one correlations.